- Email: [email protected]
Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O
TSF-35067; No of Pages 6 Thin Solid Films xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O Junghwan Kim a,⁎, Norihiko Miyokawa a, Takumi Sekiya a, Keisuke Ide a, Yoshitake Toda b, Hidenori Hiramatsu a,b, Hideo Hosono a,b, Toshio Kamiya a,b a b
Materials and Structures Laboratory, Tokyo Institute of Technology, Mailbox R3-4, 4259 Nagatsuta, Midori-ku, Yokohama, Japan Materials Research Center for Element Strategy, Tokyo Institute of Technology, Mailbox SE-6, 4259 Nagatsuta, Midori-ku, Yokohama, Japan
a r t i c l e
i n f o
Article history: Received 30 November 2015 Received in revised form 27 February 2016 Accepted 1 March 2016 Available online xxxx Keywords: Amorphous oxide semiconductor Band alignment Ionization potential Electron traps
a b s t r a c t We fabricated amorphous oxide semiconductor films, a-(Ga1–xZnx)Oy, at room temperature on glass, which have widely tunable band gaps (Eg) ranging from 3.47–4.12 eV. The highest electron Hall mobility ~7 cm2 V−1 s−1 was obtained for Eg = ~3.8 eV. Ultraviolet photoemission spectroscopy revealed that the increase in Eg with increasing the Ga content comes mostly from the deepening of the valence band maximum level while the conduction band minimum level remains almost unchanged. These characteristics are explained by their electronic structures. As these films can be fabricated at room temperature on plastic, this achievement extends the applications of flexible electronics to opto-electronic integrated circuits associated with deep ultraviolet region. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Amorphous oxide semiconductor (AOS) thin-film transistor was firstly reported in 2004 [1] and has attracted much attention due to their promising properties such as high transparency, high electron mobility, and low-temperature process compatibility. In ionic oxide materials like AOS, conduction band (CB) minimum (CBM) is formed mainly by unoccupied metal ns orbitals (n = principal quantum number), while fully-occupied O 2p orbitals contribute mainly to the valence band maximum (VBM) [2–4]. Therefore, large overlap between the ns-orbitals of next neighboring cations intervening an oxygen leads to large CB dispersion and high electron mobility even in an amorphous structure as demonstrated for amorphous In–Ga–Zn–O (a-IGZO) [1,2]. However, these materials including a-IGZO suffer from instability against visible light illumination although their band gaps (Eg) are typically larger than 3.0 eV. It is due to defect states just above VBM (nearVBM states), [5,6] which cause subgap optical responses from the photon energy ~2.4 eV [7]. To overcome this issue, we here propose ultrawide band gap AOS with significantly large band gaps ≫3.0 eV, where even subgap states do not cause the visible light absorption. To date, a variety of crystalline transparent conductive oxides (TCOs) have been reported; however, most of them contain the representative TCOs of ZnO, In2O3, SnO2, and Ga2O3 as major constituents. Development of AOSs has followed a similar manner, leading to a-In–Zn–O, a-Zn–Sn–O, a-In–Ga–O, a-In–Sn–Zn–O, a-Sn–Ga–Zn–O, a-Hf–In–Zn–O, a-Ga–Cd–O [1,8–11], etc. Among the crystalline TCOs, pure β-Ga2O3 has the largest ⁎ Corresponding author. E-mail address: [email protected] (J. Kim).
Eg of 4.9 eV [12]; however, electron conduction in pure amorphous Ga2O3 (a-Ga2O3) has not yet been reported. In this work, we investigated amorphous thin films in the (Ga 1–x Zn x )O y system, and obtained semiconducting amorphous films from x = 0 to 0.65, where Eg ranged from 3.47–4.12 eV and the highest mobility of ~7 cm2 V−1 s−1was obtained for Eg = 3.8 eV. 2. Experimental details All thin films were fabricated by pulsed laser deposition using a KrF excimer laser (wavelength: 248 nm) in an O2 gas flow at room temperature on silica glass substrates. We synthesized polycrystalline ceramic targets with four different chemical compositions, ZnO, (Ga0.35Zn0.65)Oy, (Ga0.7Zn0.3)Oy, and Ga2O3. The targets were prepared from powders of ZnO (purity, 99.999%) and Ga2O3 (purity, 99.99%), which were sintered at 1500 °C for 5 h in air. Oxygen pressure (PO2) during deposition and laser fluence were varied with 1–4 Pa and 1–4.5 J/cm2, respectively. Pulse repetition was fixed at 10 Hz. Some films were subjected to post-disposition thermal annealing at Ta = 200–600 °C in vacuum. Film structures, densities, and thickness were characterized and determined by high-resolution X-ray diffraction (HR-XRD) and grazingincidence X-ray reflectivity (GIXRR), respectively. Optical band gap values were estimated by Tauc plots for amorphous films while that of crystalline ZnO was taken from a literature. Electrical properties were measured by Hall effect with the van der Pauw configuration. Desorption of constituents was measured by thermal desorption spectroscopy (TDS). Chemical composition was analyzed by X-ray fluorescence (XRF) spectroscopy, where the chemical compositions were
http://dx.doi.org/10.1016/j.tsf.2016.03.003 0040-6090/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
2
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
calibrated using the results of inductively coupled plasma atomic emission spectroscopy. The energy levels of CBM and VBM were measured by ultraviolet photoemission spectroscopy (UPS, excited by He I and II light sources). To prepare chemically pure surfaces, Ar ion sputtering was conducted for 1 h at an acceleration voltage of 1 kV in a vacuum chamber connected to the UPS measurement chamber. Work function was determined from the cut-off energy of secondary electrons, and ionization potential (Ip) was estimated by combining the work function and the measured VBM level [13]. Electron affinity (χ, i.e., CBM level) was speculated using the band gap by χ = Ip – Eg. 3. Results and discussion 3.1. Deposition condition to obtain amorphous (Ga0.35Zn0.65)Oy We first investigated effects of laser fluence and Ta on (Ga0.35Zn0.65)Oy thin films. As shown in Fig. 1(a), the thin film deposited at a low laser fluence (1.0 J/cm2) exhibited only broad halos centered at ~35° and ~63° (the halo around 22° comes from the silica glass substrate), which indicates that the film is an amorphous phase. As the laser fluence increased, a sharp peak appeared at 32.62°, and the halos became sharper, suggesting formation and growth of crystallites. The full width at half maximum value of the halo at 32.2° for the highest laser fluence (4.5 J/cm2) film was 1.93°, which corresponds to the crystallite size of 4.5 nm by the Scherrer equation. On the other hand, this peak position
(32.2°) differs significantly from those of crystalline ZnO and β-Ga2O3; ZnO, 0002ZnO, and 111β ‐ Ga O difi.e., at 31.8°, 34.4°, and 33.5° for 1010 2 3 fractions, respectively. Thus, it is possible that nanocrystals of solidsolution (Ga,Zn)O were deposited in these films at high laser fluences. Then, we measured their chemical compositions and film densities using XRF and GIXRR, respectively. As summarized in Table 1, the film density was low, 5.27 g/cm3, for the low laser fluence of 1 J/cm2, but increased to 5.69 g/cm3 with increasing laser fluence to 4.5 J/cm2. On the other hand, the chemical composition ratio Ga:Zn was almost unchanged at ~65:35. TDS spectra for H2O (Fig. 1(c)) show that the low-density film deposited at the low laser fluence (1.0 J/cm2) exhibited high-density desorption of H2O molecules from 200–330 °C, while, the high-density film deposited at 4.5 J/cm2 exhibited much less H2O desorption. It is known that high-density impurity hydrogen N1020 cm−3 is contained in usual AOS such as a-IGZO deposited by conventional pulsed laser deposition and sputtering, which comes from the residual gas in the deposition chamber and/or supplied gases [14]. These results suggest that the low densities of the lower laser fluence films are related partly to the incorporation of more H-related impurities such as H, H2O, and OH. The incorporation of the impurities is enhanced at a lower deposition rate because the incorporation rate of hydrogen (i.e., number of hydrogen atoms per second) should be constant and determined by their densities in the residual/supplied gases, while a lower deposition rate requires a longer deposition time and consequently incorporate
Fig. 1. Structures and optical properties of (Ga0.35Zn0.65)Oy thin films. XRD patterns of (a) as-deposited thin films deposited at laser fluences between 1.0 and 4.5 J/cm2 and (b) annealed ZnO, 0002ZnO, and 111β‐Ga O ). (c) TDS spectra for m/z = 18 (H2O) thin films, deposited at laser fluence 1.0 J/cm2, at Ta = 200–600 °C. (Vertical bars show diffraction peak positions for 1010 2 3 of thin films deposited at laser fluences of 1.0 and 4.5 J/cm2. (d) Optical absorption spectra of thin films deposited at fluences of 1.0 and 4.5 J/cm2 and PO2 = 1–4 Pa. That of a polycrystalline ZnO thin film is shown for comparison.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
3
Table 1 Relationship among deposition condition, film density, and chemical composition ratio for (Ga0.35Zn0.65)Oy thin films. Laser fluence (J/cm2)
Pulse repetition (Hz)
Deposition rate (nm/min)
Film thickness (nm)
Film density (g/cm3)
Zn:Ga
1.0 1.5 3.0 4.5
10 10 10 10
7.2 9.8 26.3 34.8
107.1 110.2 115.4 117.1
5.27 5.33 5.50 5.69
64.7:35.3 64.5:35.5 64.3:35.7 65.2:34.8
more impurities. Further, as it is reported that incorporation of H/H2O stabilizes amorphous phases for (In,Sn)2O3 [15,16], In2O3 [17], etc., it is considered that the amorphous phase of the film deposited at 1 J/cm2 would be stabilized also by the impurity hydrogen. We next conducted post-deposition thermal annealing for the amorphous (Ga0.35Zn0.65)Oy thin films deposited at 1.0 J/cm2. As shown in Fig. 1(b), no distinct crystallization was observed up to 400 °C. While, a new halo feature appears at ~ 34.8° and its intensity started increasing from 500 °C. As the evolution of this halo is similar to that of the sharp peak detected in the 1.5–4.5 J/cm2 films in Fig. 1(a), it is considered that the crystallization starts from a crystallization temperature (TX) somewhere between 400 and 500 °C. As seen in Fig. 1(c), the H2O desorption started from ~200 °C and depleted up to ~330 °C. This H2O depletion temperature is obviously lower than the TX, indicating that the crystallization in this (Ga1–xZnx)Oy system requires extra thermal energy other than the desorption of impurity hydrogen that would stabilize the amorphous phase in part. In Fig. 1(d), we compared optical properties of thin films deposited with varied laser fluences. Polycrystalline- (poly-) (Ga0.35Zn0.65)Oy thin films deposited at 4.5 J/cm2 exhibited higher Urbach tail-like absorption, compared to the amorphous- (a-) (Ga0.35Zn0.65)Oy deposited at 1 J/cm2 . The tail-like absorption down to ~ 3.5 eV in the poly(Ga 0.35 Zn0.65 )Oy thin films did not change with PO2 that varied from 1 to 4 Pa, indicating that the tail-like absorption came not from oxygen deficiency but from the inherent structural property of the polycrystalline film such as nanocrystals and grain boundaries.
Next, we compared electrical properties of a-(Ga0.35Zn0.65)Oy and poly-(Ga0.35Zn0.65)Oy. As seen in Fig. 1(a), the poly-(Ga0.35Zn0.65)Oy thin film was deposited at the high laser fluence of 4.5 J/cm2, while the a-(Ga0.35Zn0.65)Oy film was deposited at 1.0 J/cm2 (the other parameters such as the target–substrate distance of 40 mm and PO2 of 4 Pa were the same for both cases). Further, both of them were annealed at 200 °C in vacuum, which improved their electrical properties due partly to reduction of excess/weakly-bonded oxygen [18]. As shown in Fig. 2(a), the Hall mobility (μe) at 300 K for the a-(Ga0.35Zn0.65)Oy (6 cm2 V−1 s−1) is doubly larger than that of the poly-(Ga0.35Zn0.65)Oy (3 cm2 V− 1 s− 1). The activation energy of μe is larger for the poly(Ga 0.35 Zn0.65 )O y (45 meV) than that for the a-(Ga 0.35 Zn0.65)O y (29 meV). The prefactors of the Arrhenius plots were very similar, μ0 = ~ 18 and ~ 17 cm 2 V− 1 s− 1 for the a-(Ga0.35 Zn0.65)Oy and the poly-(Ga0.35Zn0.65)Oy thin films, respectively. These results indicate that the extended mobility (μ0, i.e., mobility in single-crystallite or in single-domain within potential barriers) is not affected by the amorphous structure, which is similar to other AOSs including a-IGZO [1,19], and the Hall mobilities are determined by the electron transport potential barriers in CB (probably due to grain boundaries for the polycrystalline film and disordered structure for the amorphous film as reported in [19]). The carrier densities (Ne) at 300 K were ~ 2 × 1017 cm− 3 and ~ 4 × 10 17 cm − 3 for a-(Ga 0.35Zn0.65 )O y and poly-(Ga0.35Zn0.65)Oy, respectively (Fig. 2(b)). Their activation energies were very similar (~43 meV), indicating that the donor levels are located at ~0.09 eV from the CBM. 3.2. Effects of chemical composition on a-(Ga1 − xZnx)Oy Further, we investigated Ga-rich region in the a-(Ga1 − xZnx)Oy system. Fig. 3 shows XRD patterns of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga 2 O3 thin films. Only halos from the a-(Ga1 − xZnx)Oy films and the silica glass substrates were observed, substantiating that all the films were amorphous. Fig. 4 compares optical properties for different Ga contents, from which Eg values
Fig. 2. Temperature dependences of (a) μe and (b) ne for a-(Ga0.35Zn0.65)Oy and poly(Ga0.35Zn0.65)Oy thin films.
Fig. 3. XRD patterns of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3 thin films.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
4
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
Fig. 5. Band alignment of crystalline ZnO [10,21], a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3. Those for crystalline β-Ga2O3 [20], SnO2 [21], In2O3 [21], and a-IGZO [22] are also shown for comparison, where the pink regions in Ip and the green regions in χ show the variation of reported values possibly caused by fabrication/treatment conditions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 4. Optical absorption spectra of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3 thin films plotted in (a) logarithmic scale and (b) Tauc plot.
were extracted from Tauc plots (Fig. 4(b)) (note that Eg = 3.37 eV of ZnO is taken from [23] because Tauc plot is not applied to a crystalline material). Eg increased with the increase in the Ga content from 3.37 eV for crystalline ZnO, 3.47 eV for a-(Ga0.35Zn0.65)Oy, 3.78 eV for a(Ga0.70Zn0.30)Oy, and to 4.12 eV for a-Ga2O3. To investigate how electronic structure changes by the incorporation of Ga, we measured energy levels of VBM by UPS and speculated CBM using Eg, which are compared in Fig. 5. It is seen that the change in Eg results mostly from the change in IP (corresponding to the VBM level). On the other hand, the CBM levels are almost unchanged from pure ZnO to pure Ga2O3, implying that their lowest unoccupied levels and the CB widths are similar. Here, we like to discuss why the a-Ga2O3 has the deepest VBM level (i.e., the largest IP). There are several origins that affect the VBM and the CBM level which are as follows: (i) It is known that the electronic levels in cations are raised by negative Madelung potential (MP) formed by surrounding anions and vice versa, which forms an open and large band gap in ionic materials [24]; therefore, larger MP (e.g. by shorter interionic distances and larger ion charges) deepens the VBM level and raises the CBM level. Shortening interionic distances would have two effects. (ii) One is to increase the overlap of relevant atomic orbitals between adjacent cations and anions. It increases the energy splitting between the bonding orbitals and the anti-bonding ones, which deepens
the VBM level and raises the CBM level. (iii) On the other hand, it would have an opposite effect; shorter interionic distance increases the overlap between ion orbitals in the CB-constituting cations and in the VBconstituting oxygens, and increases the CB and VB band dispersions and widths, which raises the VBM level and deepens the CBM level. (iv) It is also pointed that the interaction between anion p and cation d orbitals is important; i.e., the VBM of II–VI compound semiconductor is raised by strong anti-bonding coupling between anion p and cation d orbitals [25]. The data for crystalline oxides in Fig. 5 indicates that the VBM level becomes deeper in the order of ZnO, In2O3, SnO2, and β-Ga2O3 although the VBM values have some distribution in reported values (the distribution ranges are shown by the pink regions). It is not explained simply by the ion charges because the VBM level of β-Ga3+2O2−3 is ~1 eV deeper than that of In3+2O2−3 although their formal ion charges are the same. As the Ga–O distances in β-Ga2O3 (0.180–0.208 nm, ICSD #34243) are smaller than the In–O ones in In2O3 (0.211–0.221 nm, ICSD #14387), it would be reasonable that the MP in In2O3 is smaller and the VBM is shallower than those in β-Ga2O3; however, the VBM level of SnO2 is shallower than that of β-Ga2O3 although the MP of Sn4+ O2−2 should be larger than that of β-Ga3+2O2−3. If we consider the p–d coupling effect of the above origin (iv), it can explain the order of the VBM levels from ZnO, In2O3, to [β-Ga2O3, SnO2] because the d orbital levels deepen and become apart from O 2p in the order of Zn 3d, In 4d, Ga 3d, and Sn 4d [26]. It is, however, difficult to explain the similar VBM levels between β-Ga2O3 and SnO2. The shorter Ga–O distances in β-Ga2O3 (0.180–0.208 nm) than the Sn–O ones in SnO2 (0.205–0.206 nm, ICSD #16635) would contribute to deepening the VBM level of β-Ga2O3. Finally, we should say that the VBM levels of these oxides would be affected by the above several factors, and the relatively deep Ga 3d level and the short Ga–O distances form the deep VBM compared to ZnO and In2O3. The VBM level of aGa2O3 is shallower than that of β-Ga2O3, which would be explained by the fact that the disordered structure in a-Ga2O3 causes more distribution in the O 2p levels and raises the VBM level, as reported for a-IGZO [27]. Fig. 6 compares carrier transport between the Zn-rich a(Ga 0.35 Zn0.65 )O y and the Ga-rich a-(Ga 0.70Zn0.30)O y films. The μe values at 300 K are similar: 6 cm2 V−1 s−1 and 7 cm2 V−1 s−1 for the
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
5
room temperature and have flexible tunability of the band gap and the VBM level, they would be applicable to design hole blocking layers in flexible optoelectronic devices [30,31]. Acknowledgments This work was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) though the Element Strategy Initiative to Form Core Research Center. J. Kim was supported by ACCEL Project sponsored by the Japan Science and Technology Agency (JST). H. Hiramatsu was also supported by the Japan Society for the Promotion of Science (JSPS) through a Grant-in-Aid for Scientific Research on Innovative Areas “Nano Informatics” Grant Number 25106007 and Support for Tokyotech Advanced Research (STAR). References
Fig. 6. Temperature dependences of (a) μe and (b) ne for a-(Ga0.35Zn0.65)Oy and (Ga0.70Zn0.30)Oy thin films.
a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy films, respectively (Fig. 6(a)), while their carrier densities are largely different, i.e., ~ 2 × 1017 cm− 3 and ~ 4 × 1015 cm− 3 for the a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy, respectively. As the prefactors of the Arrhenius plot (n0) are close to each other (n0 = 1.04 × 1018 and 1.49 × 1018 cm−3 for the a-(Ga0.35Zn0.65)Oy and a-(Ga0.70Zn0.30)Oy, respectively), the total donor densities would be similar. This difference is, therefore, associated with different donor levels (ED) because ED of the a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy is largely different, 0.08 and 0.32 eV, respectively (estimated from the Ea values of 0.04 and 0.16 eV by ED = 2Ea). In this (Ga1–xZnx)Oy system, plausible donor sources would be oxygen deficiency and hydrogen [28]. As the CBM level is almost the same for these materials as seen in Fig. 5, the different ED indicate that the ionization levels of these donors depend largely on the chemical composition and/or local coordination structure of the donor; in other words, the universal donor level as reported for hydrogen donor [29] would not be applied to the (Ga1–xZnx)Oy system. As seen in Fig. 5, the difference in the VBM levels is 0.28 eV and close to the difference in ED (0.24 eV); it appears that ED is more likely aligned to the VBM level. 4. Summary We deposited amorphous (Ga1 − xZnx)Oy films and discussed their structures, electron transport properties, electronic structures, and donor levels. We found that low laser fluences and low deposition rates are required to form amorphous films. The largely tuned band gap from 3.47 to 4.12 eV for x = 0.65–0 is determined almost solely by the VBM level. It was also found that the donor levels from the vacuum level are largely dependent on the Ga content and appear to align better to the VBM levels. The large band gap of Ga-rich a-(Ga1–xZnx)Oy would be applied to deep ultraviolet optoelectronic/photonic devices. This feature would also be interesting for thin-film transistors and other electronic devices that do not show light-induced degradation/light response in the visible-to-ultraviolet region. As a-(Ga1–xZnx)Oy films are fabricated at
[1] K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, H. Hosono, Room-temperature fabrication of transparent flexible thin-film transistors using amorphous oxide semiconductors, Nature 432 (2004) 488. [2] T. Kamiya, H. Hosono, Material characteristics and applications of transparent amorphous oxide semiconductors, NPG Asia Mater. 2 (2010) 15. [3] H. Hosono, Ionic amorphous oxide semiconductors: material design, carrier transport, and device application, J. Non Cryst. Solids 352 (2006) 851. [4] J. Kim, H. Hiramatsu, H. Hosono, T. Kamiya, Effects of sulfur substitution in amorphous InGaZnO4: optical properties and first-principles calculations, J. Ceram. Soc. Jpn. 123 (2015) 537. [5] K. Nomura, T. Kamiya, H. Yanagi, E. Ikenaga, K. Yang, K. Kobayashi, M. Hirano, H. Hosono, Subgap states in transparent amorphous oxide semiconductor, In–Ga– Zn–O, observed by bulk sensitive X-ray photoelectron spectroscopy, Appl. Phys. Lett. 92 (2008) 202117. [6] H. Noh, K. Chang, B. Ryu, W. Lee, Electronic structure if oxygen-vacancy defects in amorphous In–Ga–Zn–O semiconductors, Phys. Rev. B 84 (2011) 115205. [7] T. Kamiya, H. Hosono, Roles of hydrogen in amorphous oxide semiconductor, ECS Trans. 54 (2013) 103. [8] H. Chiang, J. Wager, R. Hoffman, J. Jeong, D. Keszler, High mobility transparent thinfilm transistors with amorphous zinc tin oxide channel layer, Appl. Phys. Lett. 86 (2005) 013503. [9] D. Kim, H. Kim, J. Kwon, G. Lee, H. Lee, S. Im, Sputter-deposited Ga–Sn–Zn–O thin films for transparent thin film transistors, Phys. Status Solidi A 208 (2011) 2934. [10] J. Park, T. Kim, K. Son, K. Lee, W. Maeng, H. Kim, E. Kim, K. Park, J. seon, W. Choi, M. Ryu, S. Lee, The influence of SiOx and SiNx passivation on the negative bias stability of Hf–In–Zn–O thin film transistors under illumination, Appl. Phys. Lett. 96 (2010) 262109. [11] H. Yanagi, C. Sato, Y. Kimura, I. Suzuki, T. Omata, T. Kamiya, H. Hosono, Widely bandgap tunable amorphous Cd–Ga–O oxide semiconductors exhibiting electron mobilities ≥ 10 cm2 V−1 s−1, Appl. Phys. Lett. 106 (2015) 082106. [12] M. Mohamed, C. Janowitz, I. Unger, R. Manzke, Z. Galazka, R. Uecker, R. Fornari, J. Weber, J. Varley, C.G. Van de Walle, The electronic structure of β-Ga2O3, Appl. Phys. Lett. 97 (2010) 211903. [13] H. Yanagi, K. Kim, I. Koizumi, M. Kikuchi, H. Hiramatsu, M. Miyakawa, T. Kamiya, M. Hirano, H. Hosono, Low threshold voltage and carrier injection properties of inverted organic light-emitting diodes with [Ca24Al28O64]4+(4e−) cathode and Cu2−xSe anode, J. Phys. Chem. C 113 (2009) 18379. [14] T. Miyase, K. Watanabe, I. Sakaguchi, N. Ohashi, K. Domen, K. Nomura, H. Hiramatsu, H. Kumomi, H. Hosono, Roles of hydrogen in amorphous oxide semiconductor In– Ga–Zn–O: comparison of conventional and ultra-high-vacuum sputtering, ECS J. Solid State Sci. Technol. 3 (2014) Q3085. [15] S. Ishibashi, Y. Higuchil, Y. Ota, K. Nakamura, Low resistivity indium–tin oxide transparent conductive films. I. Effect of introducing H2O gas or H2 gas during direct current magnetron sputtering, J. Vac. Sci. Technol. A A 8 (1990) 1399. [16] E. Nishimura, H. Ohkawa, P. Song, Y. Shigesato, Microstructures of ITO films deposited by d.c. magnetron sputtering with H2O introduction, Thin Solid Films 445 (2003) 235. [17] T. Koida, H. Fujiwara, M. Kondo, Hydrogen-doped In2O3 as high-mobility transparent conductive oxide, Jpn. J. Appl. Phys. 46 (2007) L685. [18] K. Ide, Y. Kikuchi, K. Nomura, M. Kimura, T. Kamiya, H. Hosono, Effects of excess oxygen on operation characteristics of amorphous In–Ga–Zn–O thin-film transistors, Appl. Phys. Lett. 99 (2011) 093507. [19] T. Kamiya, K. Nomura, H. Hosono, Origin of definite Hall voltage and positive slope in mobility-donor density relation in disordered oxide semiconductors, Appl. Phys. Lett. 96 (2010) 122103. [20] M. Mohamed, K. Irmscher, C. Janowitz, Z. Galazka, R. Manzke, R. Fornari, Schottky barrier height of Au on the transparent semiconducting oxide β-Ga2O3, Appl. Phys. Lett. 101 (2012) 132106. [21] A. Klein, C. Korber, A. Wachau, F. Sauberlich, Y. Gassenbauer, S. Harvey, D. Proffit, T. Mason, Transparent conducting oxides for photovoltaics: manipulation of Fermi level, work function and energy band alignment, Materials 3 (2010) 4892. [22] K. Lee, K. Nomura, H. Yanagi, T. kamiya, E. ikenaga, T. Sugiyama, K. Kobayashi, H. Hosono, Band alignment of InGaZnO4/Si interface by hard X-ray photoelectron spectroscopy, J. Appl. Phys. 112 (2012) 033713.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
6
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
[23] Y. Chen, D. Bagnall, H. Koh, K. Park, K. Hiraga, Z. Zhu, T. Yao, Plasma assisted molecular beam epitaxy of ZnO on c-plane sapphire: growth and characterization, J. Appl. Phys. 84 (1998) 3912. [24] P.A. Cox, Chap. 2 Models of electronic structure, Transition Metal Oxides, Oxford University Press, Oxford, 1992. [25] S. Wei, A. Zunger, Role of metal d states in II–VI semiconductors, Phys. Rev. B 37 (1988) 8958. [26] T. Kamiya, M. Kawasaki, ZnO-based semiconductors as building blocks for active devices, MRS Bull. 33 (2008) 1061. [27] W. Körner, D.F. Urban, C. Elsässer, Origin of subgap states in amorphous In–Ga–Zn– O, J. Appl. Phys. 114 (2013) 163704.
[28] C.G. Van de Walle, Hydrogen as a cause of doping in zinc oxide, Phys. Rev. Lett. 85 (2000) 1012. [29] C.G. Van de Walle, J. Neugebauer, Universal alignment of hydrogen levels in semiconductors, insulators and solutions, Nature 423 (2003) 626. [30] M. Dosmailov, L.N. Leonat, J. Patek, D. Roth, P. Bauer, M.C. Scharber, N.S. Sariciftci, J.D. Pedarning, Transparent conductive ZnO layers on polymer substrates: thin film deposition and application in organic solar cells, Thin Solid Films 591 (2015) 97–104. [31] J. Kim, N. Miyokawa, K. Ide, Y. Toda, H. Hiramatsu, H. Hosono, T. Kamiya, Room-temperature fabrication of light-emitting thin films based on amorphous oxide semiconductor, AIP Adv. 6 (2016) 015106.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O Junghwan Kim a,⁎, Norihiko Miyokawa a, Takumi Sekiya a, Keisuke Ide a, Yoshitake Toda b, Hidenori Hiramatsu a,b, Hideo Hosono a,b, Toshio Kamiya a,b a b
Materials and Structures Laboratory, Tokyo Institute of Technology, Mailbox R3-4, 4259 Nagatsuta, Midori-ku, Yokohama, Japan Materials Research Center for Element Strategy, Tokyo Institute of Technology, Mailbox SE-6, 4259 Nagatsuta, Midori-ku, Yokohama, Japan
a r t i c l e
i n f o
Article history: Received 30 November 2015 Received in revised form 27 February 2016 Accepted 1 March 2016 Available online xxxx Keywords: Amorphous oxide semiconductor Band alignment Ionization potential Electron traps
a b s t r a c t We fabricated amorphous oxide semiconductor films, a-(Ga1–xZnx)Oy, at room temperature on glass, which have widely tunable band gaps (Eg) ranging from 3.47–4.12 eV. The highest electron Hall mobility ~7 cm2 V−1 s−1 was obtained for Eg = ~3.8 eV. Ultraviolet photoemission spectroscopy revealed that the increase in Eg with increasing the Ga content comes mostly from the deepening of the valence band maximum level while the conduction band minimum level remains almost unchanged. These characteristics are explained by their electronic structures. As these films can be fabricated at room temperature on plastic, this achievement extends the applications of flexible electronics to opto-electronic integrated circuits associated with deep ultraviolet region. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Amorphous oxide semiconductor (AOS) thin-film transistor was firstly reported in 2004 [1] and has attracted much attention due to their promising properties such as high transparency, high electron mobility, and low-temperature process compatibility. In ionic oxide materials like AOS, conduction band (CB) minimum (CBM) is formed mainly by unoccupied metal ns orbitals (n = principal quantum number), while fully-occupied O 2p orbitals contribute mainly to the valence band maximum (VBM) [2–4]. Therefore, large overlap between the ns-orbitals of next neighboring cations intervening an oxygen leads to large CB dispersion and high electron mobility even in an amorphous structure as demonstrated for amorphous In–Ga–Zn–O (a-IGZO) [1,2]. However, these materials including a-IGZO suffer from instability against visible light illumination although their band gaps (Eg) are typically larger than 3.0 eV. It is due to defect states just above VBM (nearVBM states), [5,6] which cause subgap optical responses from the photon energy ~2.4 eV [7]. To overcome this issue, we here propose ultrawide band gap AOS with significantly large band gaps ≫3.0 eV, where even subgap states do not cause the visible light absorption. To date, a variety of crystalline transparent conductive oxides (TCOs) have been reported; however, most of them contain the representative TCOs of ZnO, In2O3, SnO2, and Ga2O3 as major constituents. Development of AOSs has followed a similar manner, leading to a-In–Zn–O, a-Zn–Sn–O, a-In–Ga–O, a-In–Sn–Zn–O, a-Sn–Ga–Zn–O, a-Hf–In–Zn–O, a-Ga–Cd–O [1,8–11], etc. Among the crystalline TCOs, pure β-Ga2O3 has the largest ⁎ Corresponding author. E-mail address: [email protected] (J. Kim).
Eg of 4.9 eV [12]; however, electron conduction in pure amorphous Ga2O3 (a-Ga2O3) has not yet been reported. In this work, we investigated amorphous thin films in the (Ga 1–x Zn x )O y system, and obtained semiconducting amorphous films from x = 0 to 0.65, where Eg ranged from 3.47–4.12 eV and the highest mobility of ~7 cm2 V−1 s−1was obtained for Eg = 3.8 eV. 2. Experimental details All thin films were fabricated by pulsed laser deposition using a KrF excimer laser (wavelength: 248 nm) in an O2 gas flow at room temperature on silica glass substrates. We synthesized polycrystalline ceramic targets with four different chemical compositions, ZnO, (Ga0.35Zn0.65)Oy, (Ga0.7Zn0.3)Oy, and Ga2O3. The targets were prepared from powders of ZnO (purity, 99.999%) and Ga2O3 (purity, 99.99%), which were sintered at 1500 °C for 5 h in air. Oxygen pressure (PO2) during deposition and laser fluence were varied with 1–4 Pa and 1–4.5 J/cm2, respectively. Pulse repetition was fixed at 10 Hz. Some films were subjected to post-disposition thermal annealing at Ta = 200–600 °C in vacuum. Film structures, densities, and thickness were characterized and determined by high-resolution X-ray diffraction (HR-XRD) and grazingincidence X-ray reflectivity (GIXRR), respectively. Optical band gap values were estimated by Tauc plots for amorphous films while that of crystalline ZnO was taken from a literature. Electrical properties were measured by Hall effect with the van der Pauw configuration. Desorption of constituents was measured by thermal desorption spectroscopy (TDS). Chemical composition was analyzed by X-ray fluorescence (XRF) spectroscopy, where the chemical compositions were
http://dx.doi.org/10.1016/j.tsf.2016.03.003 0040-6090/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
2
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
calibrated using the results of inductively coupled plasma atomic emission spectroscopy. The energy levels of CBM and VBM were measured by ultraviolet photoemission spectroscopy (UPS, excited by He I and II light sources). To prepare chemically pure surfaces, Ar ion sputtering was conducted for 1 h at an acceleration voltage of 1 kV in a vacuum chamber connected to the UPS measurement chamber. Work function was determined from the cut-off energy of secondary electrons, and ionization potential (Ip) was estimated by combining the work function and the measured VBM level [13]. Electron affinity (χ, i.e., CBM level) was speculated using the band gap by χ = Ip – Eg. 3. Results and discussion 3.1. Deposition condition to obtain amorphous (Ga0.35Zn0.65)Oy We first investigated effects of laser fluence and Ta on (Ga0.35Zn0.65)Oy thin films. As shown in Fig. 1(a), the thin film deposited at a low laser fluence (1.0 J/cm2) exhibited only broad halos centered at ~35° and ~63° (the halo around 22° comes from the silica glass substrate), which indicates that the film is an amorphous phase. As the laser fluence increased, a sharp peak appeared at 32.62°, and the halos became sharper, suggesting formation and growth of crystallites. The full width at half maximum value of the halo at 32.2° for the highest laser fluence (4.5 J/cm2) film was 1.93°, which corresponds to the crystallite size of 4.5 nm by the Scherrer equation. On the other hand, this peak position
(32.2°) differs significantly from those of crystalline ZnO and β-Ga2O3; ZnO, 0002ZnO, and 111β ‐ Ga O difi.e., at 31.8°, 34.4°, and 33.5° for 1010 2 3 fractions, respectively. Thus, it is possible that nanocrystals of solidsolution (Ga,Zn)O were deposited in these films at high laser fluences. Then, we measured their chemical compositions and film densities using XRF and GIXRR, respectively. As summarized in Table 1, the film density was low, 5.27 g/cm3, for the low laser fluence of 1 J/cm2, but increased to 5.69 g/cm3 with increasing laser fluence to 4.5 J/cm2. On the other hand, the chemical composition ratio Ga:Zn was almost unchanged at ~65:35. TDS spectra for H2O (Fig. 1(c)) show that the low-density film deposited at the low laser fluence (1.0 J/cm2) exhibited high-density desorption of H2O molecules from 200–330 °C, while, the high-density film deposited at 4.5 J/cm2 exhibited much less H2O desorption. It is known that high-density impurity hydrogen N1020 cm−3 is contained in usual AOS such as a-IGZO deposited by conventional pulsed laser deposition and sputtering, which comes from the residual gas in the deposition chamber and/or supplied gases [14]. These results suggest that the low densities of the lower laser fluence films are related partly to the incorporation of more H-related impurities such as H, H2O, and OH. The incorporation of the impurities is enhanced at a lower deposition rate because the incorporation rate of hydrogen (i.e., number of hydrogen atoms per second) should be constant and determined by their densities in the residual/supplied gases, while a lower deposition rate requires a longer deposition time and consequently incorporate
Fig. 1. Structures and optical properties of (Ga0.35Zn0.65)Oy thin films. XRD patterns of (a) as-deposited thin films deposited at laser fluences between 1.0 and 4.5 J/cm2 and (b) annealed ZnO, 0002ZnO, and 111β‐Ga O ). (c) TDS spectra for m/z = 18 (H2O) thin films, deposited at laser fluence 1.0 J/cm2, at Ta = 200–600 °C. (Vertical bars show diffraction peak positions for 1010 2 3 of thin films deposited at laser fluences of 1.0 and 4.5 J/cm2. (d) Optical absorption spectra of thin films deposited at fluences of 1.0 and 4.5 J/cm2 and PO2 = 1–4 Pa. That of a polycrystalline ZnO thin film is shown for comparison.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
3
Table 1 Relationship among deposition condition, film density, and chemical composition ratio for (Ga0.35Zn0.65)Oy thin films. Laser fluence (J/cm2)
Pulse repetition (Hz)
Deposition rate (nm/min)
Film thickness (nm)
Film density (g/cm3)
Zn:Ga
1.0 1.5 3.0 4.5
10 10 10 10
7.2 9.8 26.3 34.8
107.1 110.2 115.4 117.1
5.27 5.33 5.50 5.69
64.7:35.3 64.5:35.5 64.3:35.7 65.2:34.8
more impurities. Further, as it is reported that incorporation of H/H2O stabilizes amorphous phases for (In,Sn)2O3 [15,16], In2O3 [17], etc., it is considered that the amorphous phase of the film deposited at 1 J/cm2 would be stabilized also by the impurity hydrogen. We next conducted post-deposition thermal annealing for the amorphous (Ga0.35Zn0.65)Oy thin films deposited at 1.0 J/cm2. As shown in Fig. 1(b), no distinct crystallization was observed up to 400 °C. While, a new halo feature appears at ~ 34.8° and its intensity started increasing from 500 °C. As the evolution of this halo is similar to that of the sharp peak detected in the 1.5–4.5 J/cm2 films in Fig. 1(a), it is considered that the crystallization starts from a crystallization temperature (TX) somewhere between 400 and 500 °C. As seen in Fig. 1(c), the H2O desorption started from ~200 °C and depleted up to ~330 °C. This H2O depletion temperature is obviously lower than the TX, indicating that the crystallization in this (Ga1–xZnx)Oy system requires extra thermal energy other than the desorption of impurity hydrogen that would stabilize the amorphous phase in part. In Fig. 1(d), we compared optical properties of thin films deposited with varied laser fluences. Polycrystalline- (poly-) (Ga0.35Zn0.65)Oy thin films deposited at 4.5 J/cm2 exhibited higher Urbach tail-like absorption, compared to the amorphous- (a-) (Ga0.35Zn0.65)Oy deposited at 1 J/cm2 . The tail-like absorption down to ~ 3.5 eV in the poly(Ga 0.35 Zn0.65 )Oy thin films did not change with PO2 that varied from 1 to 4 Pa, indicating that the tail-like absorption came not from oxygen deficiency but from the inherent structural property of the polycrystalline film such as nanocrystals and grain boundaries.
Next, we compared electrical properties of a-(Ga0.35Zn0.65)Oy and poly-(Ga0.35Zn0.65)Oy. As seen in Fig. 1(a), the poly-(Ga0.35Zn0.65)Oy thin film was deposited at the high laser fluence of 4.5 J/cm2, while the a-(Ga0.35Zn0.65)Oy film was deposited at 1.0 J/cm2 (the other parameters such as the target–substrate distance of 40 mm and PO2 of 4 Pa were the same for both cases). Further, both of them were annealed at 200 °C in vacuum, which improved their electrical properties due partly to reduction of excess/weakly-bonded oxygen [18]. As shown in Fig. 2(a), the Hall mobility (μe) at 300 K for the a-(Ga0.35Zn0.65)Oy (6 cm2 V−1 s−1) is doubly larger than that of the poly-(Ga0.35Zn0.65)Oy (3 cm2 V− 1 s− 1). The activation energy of μe is larger for the poly(Ga 0.35 Zn0.65 )O y (45 meV) than that for the a-(Ga 0.35 Zn0.65)O y (29 meV). The prefactors of the Arrhenius plots were very similar, μ0 = ~ 18 and ~ 17 cm 2 V− 1 s− 1 for the a-(Ga0.35 Zn0.65)Oy and the poly-(Ga0.35Zn0.65)Oy thin films, respectively. These results indicate that the extended mobility (μ0, i.e., mobility in single-crystallite or in single-domain within potential barriers) is not affected by the amorphous structure, which is similar to other AOSs including a-IGZO [1,19], and the Hall mobilities are determined by the electron transport potential barriers in CB (probably due to grain boundaries for the polycrystalline film and disordered structure for the amorphous film as reported in [19]). The carrier densities (Ne) at 300 K were ~ 2 × 1017 cm− 3 and ~ 4 × 10 17 cm − 3 for a-(Ga 0.35Zn0.65 )O y and poly-(Ga0.35Zn0.65)Oy, respectively (Fig. 2(b)). Their activation energies were very similar (~43 meV), indicating that the donor levels are located at ~0.09 eV from the CBM. 3.2. Effects of chemical composition on a-(Ga1 − xZnx)Oy Further, we investigated Ga-rich region in the a-(Ga1 − xZnx)Oy system. Fig. 3 shows XRD patterns of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga 2 O3 thin films. Only halos from the a-(Ga1 − xZnx)Oy films and the silica glass substrates were observed, substantiating that all the films were amorphous. Fig. 4 compares optical properties for different Ga contents, from which Eg values
Fig. 2. Temperature dependences of (a) μe and (b) ne for a-(Ga0.35Zn0.65)Oy and poly(Ga0.35Zn0.65)Oy thin films.
Fig. 3. XRD patterns of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3 thin films.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
4
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
Fig. 5. Band alignment of crystalline ZnO [10,21], a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3. Those for crystalline β-Ga2O3 [20], SnO2 [21], In2O3 [21], and a-IGZO [22] are also shown for comparison, where the pink regions in Ip and the green regions in χ show the variation of reported values possibly caused by fabrication/treatment conditions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 4. Optical absorption spectra of as-deposited a-(Ga0.35Zn0.65)Oy, a-(Ga0.70Zn0.30)Oy, and a-Ga2O3 thin films plotted in (a) logarithmic scale and (b) Tauc plot.
were extracted from Tauc plots (Fig. 4(b)) (note that Eg = 3.37 eV of ZnO is taken from [23] because Tauc plot is not applied to a crystalline material). Eg increased with the increase in the Ga content from 3.37 eV for crystalline ZnO, 3.47 eV for a-(Ga0.35Zn0.65)Oy, 3.78 eV for a(Ga0.70Zn0.30)Oy, and to 4.12 eV for a-Ga2O3. To investigate how electronic structure changes by the incorporation of Ga, we measured energy levels of VBM by UPS and speculated CBM using Eg, which are compared in Fig. 5. It is seen that the change in Eg results mostly from the change in IP (corresponding to the VBM level). On the other hand, the CBM levels are almost unchanged from pure ZnO to pure Ga2O3, implying that their lowest unoccupied levels and the CB widths are similar. Here, we like to discuss why the a-Ga2O3 has the deepest VBM level (i.e., the largest IP). There are several origins that affect the VBM and the CBM level which are as follows: (i) It is known that the electronic levels in cations are raised by negative Madelung potential (MP) formed by surrounding anions and vice versa, which forms an open and large band gap in ionic materials [24]; therefore, larger MP (e.g. by shorter interionic distances and larger ion charges) deepens the VBM level and raises the CBM level. Shortening interionic distances would have two effects. (ii) One is to increase the overlap of relevant atomic orbitals between adjacent cations and anions. It increases the energy splitting between the bonding orbitals and the anti-bonding ones, which deepens
the VBM level and raises the CBM level. (iii) On the other hand, it would have an opposite effect; shorter interionic distance increases the overlap between ion orbitals in the CB-constituting cations and in the VBconstituting oxygens, and increases the CB and VB band dispersions and widths, which raises the VBM level and deepens the CBM level. (iv) It is also pointed that the interaction between anion p and cation d orbitals is important; i.e., the VBM of II–VI compound semiconductor is raised by strong anti-bonding coupling between anion p and cation d orbitals [25]. The data for crystalline oxides in Fig. 5 indicates that the VBM level becomes deeper in the order of ZnO, In2O3, SnO2, and β-Ga2O3 although the VBM values have some distribution in reported values (the distribution ranges are shown by the pink regions). It is not explained simply by the ion charges because the VBM level of β-Ga3+2O2−3 is ~1 eV deeper than that of In3+2O2−3 although their formal ion charges are the same. As the Ga–O distances in β-Ga2O3 (0.180–0.208 nm, ICSD #34243) are smaller than the In–O ones in In2O3 (0.211–0.221 nm, ICSD #14387), it would be reasonable that the MP in In2O3 is smaller and the VBM is shallower than those in β-Ga2O3; however, the VBM level of SnO2 is shallower than that of β-Ga2O3 although the MP of Sn4+ O2−2 should be larger than that of β-Ga3+2O2−3. If we consider the p–d coupling effect of the above origin (iv), it can explain the order of the VBM levels from ZnO, In2O3, to [β-Ga2O3, SnO2] because the d orbital levels deepen and become apart from O 2p in the order of Zn 3d, In 4d, Ga 3d, and Sn 4d [26]. It is, however, difficult to explain the similar VBM levels between β-Ga2O3 and SnO2. The shorter Ga–O distances in β-Ga2O3 (0.180–0.208 nm) than the Sn–O ones in SnO2 (0.205–0.206 nm, ICSD #16635) would contribute to deepening the VBM level of β-Ga2O3. Finally, we should say that the VBM levels of these oxides would be affected by the above several factors, and the relatively deep Ga 3d level and the short Ga–O distances form the deep VBM compared to ZnO and In2O3. The VBM level of aGa2O3 is shallower than that of β-Ga2O3, which would be explained by the fact that the disordered structure in a-Ga2O3 causes more distribution in the O 2p levels and raises the VBM level, as reported for a-IGZO [27]. Fig. 6 compares carrier transport between the Zn-rich a(Ga 0.35 Zn0.65 )O y and the Ga-rich a-(Ga 0.70Zn0.30)O y films. The μe values at 300 K are similar: 6 cm2 V−1 s−1 and 7 cm2 V−1 s−1 for the
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
5
room temperature and have flexible tunability of the band gap and the VBM level, they would be applicable to design hole blocking layers in flexible optoelectronic devices [30,31]. Acknowledgments This work was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) though the Element Strategy Initiative to Form Core Research Center. J. Kim was supported by ACCEL Project sponsored by the Japan Science and Technology Agency (JST). H. Hiramatsu was also supported by the Japan Society for the Promotion of Science (JSPS) through a Grant-in-Aid for Scientific Research on Innovative Areas “Nano Informatics” Grant Number 25106007 and Support for Tokyotech Advanced Research (STAR). References
Fig. 6. Temperature dependences of (a) μe and (b) ne for a-(Ga0.35Zn0.65)Oy and (Ga0.70Zn0.30)Oy thin films.
a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy films, respectively (Fig. 6(a)), while their carrier densities are largely different, i.e., ~ 2 × 1017 cm− 3 and ~ 4 × 1015 cm− 3 for the a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy, respectively. As the prefactors of the Arrhenius plot (n0) are close to each other (n0 = 1.04 × 1018 and 1.49 × 1018 cm−3 for the a-(Ga0.35Zn0.65)Oy and a-(Ga0.70Zn0.30)Oy, respectively), the total donor densities would be similar. This difference is, therefore, associated with different donor levels (ED) because ED of the a-(Ga0.35Zn0.65)Oy and the a-(Ga0.70Zn0.30)Oy is largely different, 0.08 and 0.32 eV, respectively (estimated from the Ea values of 0.04 and 0.16 eV by ED = 2Ea). In this (Ga1–xZnx)Oy system, plausible donor sources would be oxygen deficiency and hydrogen [28]. As the CBM level is almost the same for these materials as seen in Fig. 5, the different ED indicate that the ionization levels of these donors depend largely on the chemical composition and/or local coordination structure of the donor; in other words, the universal donor level as reported for hydrogen donor [29] would not be applied to the (Ga1–xZnx)Oy system. As seen in Fig. 5, the difference in the VBM levels is 0.28 eV and close to the difference in ED (0.24 eV); it appears that ED is more likely aligned to the VBM level. 4. Summary We deposited amorphous (Ga1 − xZnx)Oy films and discussed their structures, electron transport properties, electronic structures, and donor levels. We found that low laser fluences and low deposition rates are required to form amorphous films. The largely tuned band gap from 3.47 to 4.12 eV for x = 0.65–0 is determined almost solely by the VBM level. It was also found that the donor levels from the vacuum level are largely dependent on the Ga content and appear to align better to the VBM levels. The large band gap of Ga-rich a-(Ga1–xZnx)Oy would be applied to deep ultraviolet optoelectronic/photonic devices. This feature would also be interesting for thin-film transistors and other electronic devices that do not show light-induced degradation/light response in the visible-to-ultraviolet region. As a-(Ga1–xZnx)Oy films are fabricated at
[1] K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, H. Hosono, Room-temperature fabrication of transparent flexible thin-film transistors using amorphous oxide semiconductors, Nature 432 (2004) 488. [2] T. Kamiya, H. Hosono, Material characteristics and applications of transparent amorphous oxide semiconductors, NPG Asia Mater. 2 (2010) 15. [3] H. Hosono, Ionic amorphous oxide semiconductors: material design, carrier transport, and device application, J. Non Cryst. Solids 352 (2006) 851. [4] J. Kim, H. Hiramatsu, H. Hosono, T. Kamiya, Effects of sulfur substitution in amorphous InGaZnO4: optical properties and first-principles calculations, J. Ceram. Soc. Jpn. 123 (2015) 537. [5] K. Nomura, T. Kamiya, H. Yanagi, E. Ikenaga, K. Yang, K. Kobayashi, M. Hirano, H. Hosono, Subgap states in transparent amorphous oxide semiconductor, In–Ga– Zn–O, observed by bulk sensitive X-ray photoelectron spectroscopy, Appl. Phys. Lett. 92 (2008) 202117. [6] H. Noh, K. Chang, B. Ryu, W. Lee, Electronic structure if oxygen-vacancy defects in amorphous In–Ga–Zn–O semiconductors, Phys. Rev. B 84 (2011) 115205. [7] T. Kamiya, H. Hosono, Roles of hydrogen in amorphous oxide semiconductor, ECS Trans. 54 (2013) 103. [8] H. Chiang, J. Wager, R. Hoffman, J. Jeong, D. Keszler, High mobility transparent thinfilm transistors with amorphous zinc tin oxide channel layer, Appl. Phys. Lett. 86 (2005) 013503. [9] D. Kim, H. Kim, J. Kwon, G. Lee, H. Lee, S. Im, Sputter-deposited Ga–Sn–Zn–O thin films for transparent thin film transistors, Phys. Status Solidi A 208 (2011) 2934. [10] J. Park, T. Kim, K. Son, K. Lee, W. Maeng, H. Kim, E. Kim, K. Park, J. seon, W. Choi, M. Ryu, S. Lee, The influence of SiOx and SiNx passivation on the negative bias stability of Hf–In–Zn–O thin film transistors under illumination, Appl. Phys. Lett. 96 (2010) 262109. [11] H. Yanagi, C. Sato, Y. Kimura, I. Suzuki, T. Omata, T. Kamiya, H. Hosono, Widely bandgap tunable amorphous Cd–Ga–O oxide semiconductors exhibiting electron mobilities ≥ 10 cm2 V−1 s−1, Appl. Phys. Lett. 106 (2015) 082106. [12] M. Mohamed, C. Janowitz, I. Unger, R. Manzke, Z. Galazka, R. Uecker, R. Fornari, J. Weber, J. Varley, C.G. Van de Walle, The electronic structure of β-Ga2O3, Appl. Phys. Lett. 97 (2010) 211903. [13] H. Yanagi, K. Kim, I. Koizumi, M. Kikuchi, H. Hiramatsu, M. Miyakawa, T. Kamiya, M. Hirano, H. Hosono, Low threshold voltage and carrier injection properties of inverted organic light-emitting diodes with [Ca24Al28O64]4+(4e−) cathode and Cu2−xSe anode, J. Phys. Chem. C 113 (2009) 18379. [14] T. Miyase, K. Watanabe, I. Sakaguchi, N. Ohashi, K. Domen, K. Nomura, H. Hiramatsu, H. Kumomi, H. Hosono, Roles of hydrogen in amorphous oxide semiconductor In– Ga–Zn–O: comparison of conventional and ultra-high-vacuum sputtering, ECS J. Solid State Sci. Technol. 3 (2014) Q3085. [15] S. Ishibashi, Y. Higuchil, Y. Ota, K. Nakamura, Low resistivity indium–tin oxide transparent conductive films. I. Effect of introducing H2O gas or H2 gas during direct current magnetron sputtering, J. Vac. Sci. Technol. A A 8 (1990) 1399. [16] E. Nishimura, H. Ohkawa, P. Song, Y. Shigesato, Microstructures of ITO films deposited by d.c. magnetron sputtering with H2O introduction, Thin Solid Films 445 (2003) 235. [17] T. Koida, H. Fujiwara, M. Kondo, Hydrogen-doped In2O3 as high-mobility transparent conductive oxide, Jpn. J. Appl. Phys. 46 (2007) L685. [18] K. Ide, Y. Kikuchi, K. Nomura, M. Kimura, T. Kamiya, H. Hosono, Effects of excess oxygen on operation characteristics of amorphous In–Ga–Zn–O thin-film transistors, Appl. Phys. Lett. 99 (2011) 093507. [19] T. Kamiya, K. Nomura, H. Hosono, Origin of definite Hall voltage and positive slope in mobility-donor density relation in disordered oxide semiconductors, Appl. Phys. Lett. 96 (2010) 122103. [20] M. Mohamed, K. Irmscher, C. Janowitz, Z. Galazka, R. Manzke, R. Fornari, Schottky barrier height of Au on the transparent semiconducting oxide β-Ga2O3, Appl. Phys. Lett. 101 (2012) 132106. [21] A. Klein, C. Korber, A. Wachau, F. Sauberlich, Y. Gassenbauer, S. Harvey, D. Proffit, T. Mason, Transparent conducting oxides for photovoltaics: manipulation of Fermi level, work function and energy band alignment, Materials 3 (2010) 4892. [22] K. Lee, K. Nomura, H. Yanagi, T. kamiya, E. ikenaga, T. Sugiyama, K. Kobayashi, H. Hosono, Band alignment of InGaZnO4/Si interface by hard X-ray photoelectron spectroscopy, J. Appl. Phys. 112 (2012) 033713.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003
6
J. Kim et al. / Thin Solid Films xxx (2016) xxx–xxx
[23] Y. Chen, D. Bagnall, H. Koh, K. Park, K. Hiraga, Z. Zhu, T. Yao, Plasma assisted molecular beam epitaxy of ZnO on c-plane sapphire: growth and characterization, J. Appl. Phys. 84 (1998) 3912. [24] P.A. Cox, Chap. 2 Models of electronic structure, Transition Metal Oxides, Oxford University Press, Oxford, 1992. [25] S. Wei, A. Zunger, Role of metal d states in II–VI semiconductors, Phys. Rev. B 37 (1988) 8958. [26] T. Kamiya, M. Kawasaki, ZnO-based semiconductors as building blocks for active devices, MRS Bull. 33 (2008) 1061. [27] W. Körner, D.F. Urban, C. Elsässer, Origin of subgap states in amorphous In–Ga–Zn– O, J. Appl. Phys. 114 (2013) 163704.
[28] C.G. Van de Walle, Hydrogen as a cause of doping in zinc oxide, Phys. Rev. Lett. 85 (2000) 1012. [29] C.G. Van de Walle, J. Neugebauer, Universal alignment of hydrogen levels in semiconductors, insulators and solutions, Nature 423 (2003) 626. [30] M. Dosmailov, L.N. Leonat, J. Patek, D. Roth, P. Bauer, M.C. Scharber, N.S. Sariciftci, J.D. Pedarning, Transparent conductive ZnO layers on polymer substrates: thin film deposition and application in organic solar cells, Thin Solid Films 591 (2015) 97–104. [31] J. Kim, N. Miyokawa, K. Ide, Y. Toda, H. Hiramatsu, H. Hosono, T. Kamiya, Room-temperature fabrication of light-emitting thin films based on amorphous oxide semiconductor, AIP Adv. 6 (2016) 015106.
Please cite this article as: J. Kim, et al., Ultrawide band gap amorphous oxide semiconductor, Ga–Zn–O, Thin Solid Films (2016), http://dx.doi.org/ 10.1016/j.tsf.2016.03.003